On June 21, 2026, an undersea cable connecting Saint Lucia, in the Caribbean, to the rest of the internet was cut. According to Cloudflare's quarterly report on internet disruptions, published in July 2026, traffic on a local carrier's network, Karib Cable, fell to essentially zero for about 20 hours, and traffic for the entire country dropped by roughly 60% that week. It wasn't an attack or a system failure. It was a single physical point in the network that stopped working.
The distance between that example and a mid-sized company is smaller than it looks. The internet that reaches an office, a store, or a factory almost always comes in through a single path: one cable, one carrier, one contract. When that path fails — a cut fiber, a fault in the carrier's own equipment, roadwork that severs the line — it isn't just email that stops. It's the system that issues invoices, the card machine, the cloud software, and the phone, which today runs over the same internet connection too.
That makes this a decision for whoever runs the company, not only for whoever handles IT. A second internet connection costs one more monthly bill. A drop with no second connection costs the entire operation stopped, with no one able to say when it comes back.
What happens when the only link drops
In a company with a single link, the whole day depends entirely on a carrier it doesn't control. The invoicing system runs in the cloud, the card machine needs internet, inventory is checked through a remote panel. When the link drops, all of it stops at once — and nobody inside the company knows whether the problem is the carrier's, the router's, or a cut cable somewhere down the street.
This instability is not the exception. Anatel data compiled by Teleco shows that Brazil's fixed broadband complaint rate reached 1.11 complaints per thousand subscribers in December 2025, with specific carriers passing 5 complaints per thousand subscribers in the same month. The number moves month to month, but it never reaches zero — the connection a company depends on is exposed to a failure it didn't choose and can't prevent on its own.
Cloudflare's quarterly report for the same period shows how varied the causes behind an outage can be: an earthquake in Venezuela knocked out traffic for carriers such as Fibex Telecom, which serves 1.6 million users; a power outage in Tanzania took the country offline for at least five hours; governments in Sudan and Iraq deliberately shut down internet access during school exam periods, for up to three and a half hours at a time. The causes differ — a cut cable, a power failure, someone else's decision — but the effect is the same for anyone with only one way in: nothing works.
What changes the outcome isn't avoiding the cause — no company can prevent a cable cut down the street or a fault at the carrier. It's having a second path ready to take over when the first one stops.
Why the common approach doesn't work
The common approach is to buy whichever internet plan is cheapest, without asking which route the cable takes or who else depends on it. When it drops, the only available move is calling the carrier's support line and waiting — often without knowing how long it will take, because whoever answers on the other end doesn't know either.
Another version of the same approach is trusting that the battery backup solves it. It keeps the computer and the router running through a power cut — but if the problem is the line coming into the building, not the internal power, having equipment turned on doesn't help: there is simply no internet for it to connect to.
It's also common to have a second plan contracted "just in case" — but from the same carrier, or running along the same street on the same pole. A cable cut there takes both links down at the same time, and the company only finds out when they fail together.
What has to be in place
An environment prepared for that drop rests on concrete mechanisms, not luck.
A second link, from a different carrier and a different physical path. Contracting from the same carrier, or a cable that runs down the same street, doesn't solve it — the cut that takes one down takes the other with it. The second link needs to come from a different company and, when possible, through a different medium: fiber from one carrier and a radio or cellular link from another, for example.
An automatic switch between the two links when one drops. Without it, the second link only helps if someone notices the outage and manually swaps the cable on the equipment — which, in a store with no IT person nearby, can take hours before anyone even notices.
A list of what needs to come back first. Not everything needs the same priority the moment the link drops: the invoicing system and the card machine come before the video playing on a screen at the front desk.
One single point of contact dealing with the carrier. Someone who already knows the contract number, the line's history, and how to open a technical ticket — instead of the business owner waiting on hold while a customer waits in the store.
A record of every outage, with date, duration, and cause. Without that history, nobody knows whether the problem keeps repeating with the same carrier or was a one-off — and the carrier has no reason to act differently next time.
A cost estimate before contracting the second link. A second connection carries its own monthly cost; weighing that fixed cost against the loss from an entire afternoon offline is what makes the decision straightforward.
This is how Skills IT works: with a second link tested and ready to take over on its own when the first one drops, and a single point of contact who deals with the carrier on the client's behalf.
What the company gains
The gain from having two internet paths isn't abstract: it's the store still selling, the invoice still going out, and the phone still ringing the minute the main carrier runs into trouble.
It also shows up in what stops happening. Nobody spends the morning on hold with the carrier's support line while a customer waits in the store. Nobody loses the day guessing whether the problem is the carrier's or the equipment's own fault. And the cost of connectivity becomes predictable — two fixed monthly bills, instead of a variable loss every time the main link fails.
After a single cable was cut, traffic for the entire country dropped by about 60% compared to the previous week.
There's also a gain in decision-making: with a record of every outage, whoever decides knows whether it's worth renegotiating with the current carrier, switching providers, or whether the second link already solved the problem. That's a decision made with information, not one made in the dark in the middle of an outage. It's also why dealing with the carrier tends to move off the business owner's plate and onto whoever already does it every day, the way Skills IT does for other companies.
Questions to bring to the next meeting
Before reviewing the internet plan again, it's worth asking what exists for the day it fails:
- Does the company have a second internet connection, from a different carrier and a different physical path? If it's from the same carrier or the same pole, the single point of failure is still there.
- Does the switch between the two links happen on its own, or does someone need to notice the drop and act? If it depends on someone noticing, the downtime is however long it takes that person to show up.
- Is there a list of what needs to come back online first when the internet drops? Without one, everything competes for the same bandwidth at the exact moment it's scarcest.
- Who calls the carrier when the link drops, and does that person know the contract's history? If the answer is "whoever happens to be around," every outage starts from zero.
- Does the company know how many times the internet has dropped in the last six months, and for how long? Without that number, there's no way to tell whether the problem is getting worse or was already solved.
